Redpine raises €6.8M from NordicNinja for API of licensed data for AI agents

Redpine raised €6.8 million in a funding round led by NordicNinja to build API access to licensed data for AI agents. The company's total investment reached €9 million, and this bet looks like an attempt to close one of the narrowest gaps in the agent economy: legal access to quality data.

What Redpine builds

Redpine is not building yet another data labeling tool, but rather an infrastructure layer through which AI agents can request and pay for premium data sets in real time. Essentially, the company wants to become an intermediary between licensed content providers and programs that make decisions without human intervention. The headless API format is important here: the product is not tied to an interface for analysts or operators, but is designed for direct machine consumption, where the request, response, and payment occur within a single stream.

This is a tangible pain point for the market. An agent needs more than just the ability to reason or execute commands in a browser — it also needs reliable sources of facts, catalogs, metadata, and other structured signals that can be used without risk of violating a license. If Redpine can standardize such access, developers will have a cleaner path to commercial data on the market than collecting it piecemeal through separate contracts and custom integrations.

How the model works

Based on the description, the platform operates on a token-based model: an agent makes a request to the desired dataset and pays for usage as it is accessed. This is similar to API economics in cloud services, but adapted to the world of AI agents, where the number of requests can be very high and the value of data depends on the context of the task. This approach allows licensed content to be turned into a service that can be called programmatically rather than purchased in packages or downloaded manually.

• Access to premium licensed data without manual purchasing
• Request and payment in real time within an agent scenario
• Integration via API without a separate user interface
• A more transparent model for data providers and agent developers

Redpine's key difference from better-known players is that the company is approaching this from the API angle rather than data labeling. The article mentions Scale AI, Appen, and Defined.ai as competitors, but their basic logic historically has been built around preparing and annotating data for model training. Redpine is trying to occupy a different layer — not training, but operating agents after release. If this niche grows, those who can give machines not just a dataset, but instant and legally clean access to the needed data fragment, will win.

Why the round matters

The new round was led by NordicNinja fund, and the total amount of funds raised grew to €9 million. For an early-stage infrastructure company, this is not just a financial cushion, but a signal that investors see demand specifically for a data layer for agent products. Notably, the composition of backers stands out: the seed was backed by investors connected to OpenAI, Perplexity, and Spotify. Such a lineup does not guarantee success, but it shows that the thesis about a shortage of licensed data for agents sounds convincing even to people from major technology ecosystems. This also reflects a broader market shift.

Attention is gradually shifting from the models themselves to the infrastructure around them: observability, payments, security, rights management, and data. While some companies argue over which model handles a task better, others are building the pipes without which autonomous agents cannot work in production. Redpine clearly is betting on exactly such a layer — boring on the outside, but critically important on the inside.

What this means

If the AI agent market truly moves from demos to mass adoption, the winners will be not only model creators, but also infrastructure providers for data access. Redpine is trying to occupy a place in this stack before others: offering developers a unified API and content owners a clear way to monetize. If the model works, agent products and business will have a more legal and predictable source of quality data in their daily operations.